This Technical Report describes the steps necessary to use and to analyse two-level fractional factorial designs through illustration with six distinct applications of this methodology.
NOTE1 Each of these six illustrations is similar in that resource constraints precluded the possibility of naively running full factorial designs. Other commonalities among the six examples are noted [e.g. study objective, two levels for factors, response variable(s), factors affecting the response]. On the other hand, the individual illustrations have some salient features that are distinct.
NOTE2 The examples suggest the spectrum of possibilities both in application area and in choice of fractional factorial designs. Fractional factorial designs can be used to identify important factors for subsequent investigation (screening design) and can in some cases provide a viable understanding of the process under study. Fractional factorial designs include screening designs and designs that have been popularized by Genichi Taguchi.
NOTE3 Fractional factorial experiments are sometimes employed by individuals (so-called “black belts?? or “green belts??) associated with Six Sigma methods. Six Sigma methods are concerned with problem solving and continuous improvement. A fractional factorial experiment can be a cost-effective tool for obtaining timely improvements of processes and products. Detailed discussions and treatment of other tools employed by Six Sigma practitioners can be identified in various ISO/TC69/SC7 documents.